Statistical process control in assessing water quality in the Doce river basin after the collapse of the Fundão dam (Mariana, Brazil).

Abstract

The process of extracting information from data generated in environmental monitoring programs is often carried out using statistical tools, with Statistical Process Control (SPC) showing great potential for application in environmental monitoring. In November 2015, millions of cubic metres of tailings were dumped into the basin of the River Doce with the collapse of the Fundão dam. A study of the impact of this incident requires new approaches in data monitoring and processing, so it was sought to evaluate, using SPC tools, changes in water quality in the basin of the River Doce following the collapse of the dam. Using process charts and the process capability index (PCI), water quality parameters in the Doce and Carmo rivers were evaluated between 2009 and 2020. There, turbidity has improved since 2018, and Mn since 2016. Control charts showed that by December 2020 dissolved Fe was still not within normal pre-event fluctuation patterns. The PCI value showed that the situation worsened after the event for each of the parameters, with the lowest values for Mn and E. coli. By using a reference period, SPC makes it possible to infer the permanence of the impact of extreme pollution on the waterbody, which can be used in the routine monitoring of water quality in such events.